And fans of college football have always had a love/hate relationship with polls, too. When they are kind to your team or conference, they're great; when they're not, they're biased, unfair, or voters are stupid - or some variation on that theme.
The Sports Network recently revealed the complete list of their voters in their Top 25 poll - in an effort to provide some accountability to some of their more questionable teams making cameo appearances in the "Others Receiving Votes" area. Their revelation, however, has raised more questions than answers. (more)
In 1934, Associated Press began running their own polls of sports writers to determine whom they thought was the best college football team in the country. The methodology was very simple: 25 points for the first-placed team on everyone's ballot, 24 for the second, and so on, down to 1 point for the twenty-fifth.
The members of the FBS AP poll today, are sixty hand-picked voters who cover FBS football either nationally or regionally, broken down by state. No more than four media members from each state are picked. The full list of AP voters can be found here.
While the AP poll has not been without controversy - it was dumped from the BCS equations thanks to shenanigans over the years in order to preserve bowl payouts - it is worthy of note that the number of voters is limited, and regionally balanced.
The Sports Network's poll is comprised of a hodge-podge of different folks - employees at the Sports Network, sports information directors of specific schools, conference people, media people, and other parties. While the AP's voters total sixty, TSN's totals 151 voters.
(Full disclosure: in the past, I have been a TSN voter, but I am not a voter this year.)
That's not bad in and of itself that there are so many voters, or that a large number of opinions are expressed. But when there aren't strict rules as to how many voters there are per conference or per region, one has to wonder how it affects the end result of their poll.
*****
For example, the CAA consists of ten teams, but in their voter list (which they have attributed by conference), there are seventeen voters.
While you can't assume that each voter pumps up the team they cover (or, in many cases, the school or conference that employs them), that's 1.7 votes per school. We can reasonably assume that all those voters are extremely well versed in the CAA - and probably are less versed in, say, Patriot League teams, unless their school has played one or two of them.
(This number excludes the votes from media markets of future CAA members Georgia State, Old Dominion, which would make the numbers even larger. It also excludes the employees of TSN and the independent voters not affiliated with a conference.)
This fact is not the only reason why a 3-3 team with wins over Coastal Carolina, Elon and UMass is ranked in the Top 25 and a Patriot League team who outscored its opposition 122-16 over the last three weekends can't crack the Top 25. But I bet it's a pretty big reason.
Looking at the number of voters per region, too, it seems like things are not necessarily fair and balanced either.
Excluding the MEAC, Pioneer League and Independents, who are flung across regions, I separated the numbers of voters into regional blocks. The blocks are, loosely: Northeast (CAA, Patriot, Ivy, NEC), Midwest (MVC, OVC), South (Big South, SoCon), Gulf (Southland, SWAC), and West (Big Sky, Great West):
Northeast: 34 schools, 43 voters
Midwest: 18 schools, 30 voters
South: 16 schools, 24 voters
Gulf: 18 schools, 18 voters
West: 14 schools, 23 voters
Indy, Pioneer, MEAC: 29 schools, 31 voters
(I understand the regional breakdown is not perfect. It's Friday, and I was staying up late watching the Phillies. I'm lazy, but you still get the general idea.)
Looking at this data, pity your school in the TSN poll if you're from the Gulf region. Your school, loosely, will have 1 voter per school. Contrast to the Northeast (1.26), South (1.5) or West (1.64).
All of a sudden, it's not such a mystery why only one school from that region - Stephen F. Austin - is in the TSN Top 25.
This regional breakdown, in my opinion, matters. If you're a East Coast school, there's more likely to have been cross-conference play with other Northeast conferences - which could influence your vote totals. For example, I freely admit I voted Villanova No. 1 in the polls I vote in after seeing them in person (though I saw Appalachian State play on TV, and they are obviously very good, too).
But the poll is not broken beyond repair.
It does not explain, for example, why Appalachian State is consistently ranked No. 1 in the TSN polls while many other polls and indexes have Delaware ranked No. 1. It's not a name recognition thing: both Delaware and Appalachian State have national recognition for their football programs.
Furthermore, with 111 voters putting them No. 1 (out of 151 voters), it's not even close. No blocks of voters are controlling that - that's a clear majority. It seems like popular opinion is truly driving App State to No. 1 - and that's what the poll is supposed to track.
If you rebalanced the regions, or reduced the number of voters per conference, there's compelling evidence that they would still vote Appalachian State No. 1. Whether it's because they're on TV often, or they have better national recognition, it's hard to be sure. But it's clear that the TSN poll is getting the right team voted No. 1 by the voters - and that matters.
****
I don't bring all this up to try to say to TSN to give up on their poll. It's a hard, sometimes thankless, job to put these polls together and by doing so, I realize - maybe more than most - that by their nature polls (and indexes) are imperfect. (That's why they shouldn't be used to determine champions, like they are in FBS.)
But an inadvertent by-product of revealing their voters - to me - shows that after the season, TSN ought to look into tweaking - I'd stop short of saying reforming - their polls for next year.
Perhaps there are some other national voices that could be more of a moderating effect on the overall polls, and could have the effect of adding representation from more schools across conferences.
Perhaps they could be grouped more regionally, or given a certain number of slots per conference or school, to make sure all regions are somewhat represented equally.
I'm not suggesting that I won't be happy if Colgate or Lehigh should be getting No. 1 votes in the poll - clearly, they shouldn't be, so far, this year. Neither team has done anything to deserve that sort of consideration.
But maybe, if the methodology of the poll could be revisited after the season, perhaps next year the poll could be a better representation - geographically as well as per-conference - than it appears to be this year.
*****
Speaking of polls, the second Lambert Poll was released this week, with Colgate and Robert Morris tied for No. 9. Lehigh is receiving votes in the poll - something to watch, as Lehigh fans.
Comments